Introduction

The course will go through the basic theory of multi-way analysis and provide hands-on experience with how to analyze data in MATLAB. Detailed knowledge will be obtained of how to apply especially PARAFAC and N-PLS in different kinds of settings (spectroscopic as well as non-spectroscopic data). Some more advanced aspects of multi-way analysis will also be explained. These include the PARAFAC2 model e.g. for handling data with shifts and how to preprocess three-way data.

Analytical chemists/graduate students, designers of process analyzers, chemometricians interested in Multi-way methods.

Rasmus Bro is Professor at the Royal Veterinary and Agricultural University, Denmark. He has been teaching chemometrics and Multi-way methods at his own University and in dedicated continuing education courses in several countries. He, and his chemometrics group, is well-known also in several other topics related to chemometrics, such as, the chemometrics literature base, the N-way toolbox for Matlab, development of mathematical theories, and active research work on field of food industry as well as on various applied chemistry applications.

9:45-10:15 Check-in 

10:15-12:30 Introduction to three-way data (Theory and exercises): a) Definition of three-way and multi-way data; b) Typical types of data; c) How to handle and plot three-way data in matlab 

12:30-13:30 Lunch (provided) 

13:30-15:30 The PARAFAC model and its properties: a) The PARAFAC model; b) Basic PARAFAC fitting; c) Uniqueness; d) Number of components 

15:30-16:00 Coffee Break 

16:00-17:30 Continuing PARAFAC, exploring its properties

Get-together or dinner at city (around 20:00).

9:00-10:15 NPLS regression: a) Properties; b) Comparing two-way & three-way PLS 

10:15-10:45 Coffee Break 

10:45-12:30 Continuing NPLS 

12:30-13:30 Lunch (provided) 

13:30-14:30 PARAFAC2: a) Properties; b) How to do it in practice 

14:30-15:00 Coffee Break 

15:00-17:00 Continuing PARAFAC2

9:00-10:15 The Tucker3 model: a) Difference between Tucker3 and PARAFAC; b) The core array  

10:15-10:45 Coffee Break 

10:45-12:30 Preprocessing: Centering and scaling 

12:30-13:30 Lunch (provided) 

13:30-15:00 Handling missing data. Practical data analysis with multiway methods 

15:00-15:30 Concluding session with coffee

Fee: 550 €

Included in fee: Lectures, lunches, coffee and tea